Skip to content

Datasets

Hugging Face datasets with plain-language descriptions: what's inside and what models to train on them. 4,788 in the catalog so far, with new ones added every hour.

Data type

For which models

Find a dataset for your goal

Answered by Gemini 4 · the chat is kept only in this tab

Datasetgeronimobasso/drone-audio-detection-samples

Drone Audio Detection Samples

A large open collection of drone audio samples (16 kHz, mono) for training acoustic detection systems.

  • Audio classification
  • English
2.2K 42100K–1M rows
Datasetcapleaf/viVoice

Vietnamese multi-speaker TTS corpus

Audio and text in Vietnamese for training multi-speaker speech synthesis.

  • Text → speech
  • Vietnamese
2K 97100K–1M rows🔒 on request
DatasetMiniMaxAI/TTS-Multilingual-Test-Set

Multilingual test set for voice cloning

Speaker audio samples and text phrases in 24 languages for zero-shot speech synthesis evaluation by MiniMaxAI.

  • Text → speech
2K 47up to 1K rows
Datasetdeepvk/NonverbalTTS

Non-verbal vocalizations for TTS

17 hours of English speech with annotations of non-verbal sounds and emotions for expressive TTS.

  • Text → speech
  • English
1.9K 771K–10K rows
Datasetlinhtran92/viet_bud500

500-hour Vietnamese ASR corpus

About 500 hours of Vietnamese speech with texts: podcasts, travel, books, food, and various regional accents.

  • Automatic speech recognition
  • Vietnamese
1.9K 74100K–1M rows🔒 on request
Datasetnvidia/AudioSkills

Audio QA for expert reasoning

About 10 million audio-question-answer pairs for training audio-language models.

  • Audio text → text
  • English
1.7K 1071M–10M rows
DatasetMBZUAI/ArVoice

Multi-speaker vocalized Arabic speech corpus

MBZUAI Modern Standard Arabic audio dataset with multiple speakers and full diacritics in transcripts.

  • Text → speech
  • Text → audio
  • Automatic speech recognition
  • Arabic
1.5K 3410K–100K rows
Datasetespnet/ace-opencpop-segments

Chinese singing voice segments

A large audio-text corpus for synthesis and recognition of Chinese vocals.

  • Text → audio
  • Audio → audio
  • Automatic speech recognition
  • Chinese
1.5K 10100K–1M rows
Datasetgarystafford/deepfake-audio-detection

Balanced audio dataset for deepfake detection

1,866 English audio recordings (evenly split between real and synthesized) to train voice deepfake detectors.

  • Audio classification
  • English
1.4K 101K–10K rows
Datasetotoearth/otoSpeech-full-duplex-processed-141h

English full-duplex dialogues (141 hours)

Processed audio recordings of live conversations between two English speakers for training voice models.

  • Audio → audio
  • English
1.4K 36up to 1K rows🔒 on request
DatasetBingsu/zeroth-korean

Korean speech corpus with transcriptions

About 53 hours of Korean speech with text transcriptions for training and evaluating ASR models.

  • Automatic speech recognition
  • Korean
1.4K 4610K–100K rows
Datasets3prl/superb

Speech model evaluation benchmark (SUPERB)

English audio and text data for evaluating universal speech processing models.

  • Automatic speech recognition
  • Audio classification
  • English
1.3K 34100K–1M rows
DatasetTeraTTS/webinars

Russian Webinar Audio Recordings with Transcripts

Large Russian speech corpus from webinars (over 315 thousand fragments) with texts for ASR and TTS tasks.

  • Automatic speech recognition
  • Text → speech
  • Russian
1.3K 20100K–1M rows
Datasetleduckhai/MultiMed

Multilingual medical speech dataset

An audio dataset for medical speech recognition across multiple languages.

  • Automatic speech recognition
  • Vietnamese
  • English
  • German
  • +2
1.2K 2610K–100K rows
DatasetPorYoung/Kasugano-Sora

Kasugano Sora audio dataset

Anime character audio clips of Kasugano Sora for speech generation training.

    1.1K 81K–10K rows